SCAPULAR MUSCLE ELECTROMYOGRAPHIC ACTIVITY DURING ABDUCTION EXERCISES IN THE SCAPULAR PLANE IN THREE POSITIONS

Abstract

Background

Hyperactivity of the anterior deltoid (AD) has been shown to produce adverse effects on subacromial space width as a result of humeral head superior translation during rehabilitation exercises used with overhead athletes. Also, the importance of the ratio of upper trapezius (UT) to lower trapezius (LT) muscle activity has been examined during rehabilitation exercises particularly for those who develop scapular dyskinesis.

Hypothesis/Purpose

The purpose of this study was to investigate the level of LT and SA muscle activity during scapular plane elevation (scaption) in three positions while maintaining a moderate level of AD muscle activity. A secondary purpose was to identify the ratio of UT to LT muscle activity during the varied scaption exercises. The authors hypothesized that the activation of these two important muscles and the UT/LT ratio would vary with exercise position and throughout the range of scapular plane elevation.

Methods

Fourteen active young subjects performed scaption exercises in three different positions: standing (STAN), quadruped (QUAD), and prone (PRON) with three different weight loads: 0 kg, 1.8 kg, and 4.1 kg. Surface electromyography (EMG) was used to record muscular activity. Tested muscles included the UT, LT, SA, AD, and posterior deltoid muscles on the dominant side.

Results

QUAD scaption exercises with a load of 1.8 kg at 4 sec after the initial movement activated the LT muscle up to 49% of maximum voluntary isometric contraction (MVIC) while maintaining a moderate level of AD muscle activity (30% MVIC). STAN scaption exercises with the weight load of 1.8 kg at 3 sec after the initial movement activated 43% MVIC of the SA muscle while maintaining a moderate level of AD muscle activity (39% MVIC). The PRON condition generated significantly less SA muscle activity with both 1.8 and 4.1 kg weight loads than during the QUAD condition. The ratios of UT to LT muscle activity were significantly less in QUAD than those of STAN up to 4 sec after the initial movement. No significant difference was observed in the UT/LT ratio between QUAD and PRON conditions.

Conclusion

QUAD scaption exercise effectively activated both LT and SA muscles without over activating the AD and produced favorable ratios of UT to LT muscle activity.

Level of Evidence

Descriptive Cohort Study, Level 4

INTRODUCTION

Upward rotation of the scapula is necessary to elevate the arm, and coordinated muscular activity during scapulothoracic movement is essential to stabilize the scapula during the glenohumeral movements.¹ Muscle activation of the lower trapezius has been particularly interesting to researchers and clinicians who deal with preventive shoulder injury exercises for overhead athletes.^2-8 The importance of the balance of muscular activity has also been advocated, specifically, the ratio of the upper trapezius muscle (UT) to the lower trapezius (LT) muscle (UT/LT ratio) during scapular exercises because hyperactivity of the UT muscle impairs the quality of upward rotation of the scapula during arm elevation.⁹ Patients with subacromial pain syndrome demonstrated a significant increase in the UT/LT ratio during glenohumeral joint (GHJ) abduction in the scapular plane, also known as scaption, compared to their uninjured counterparts.¹⁰ Furthermore, in patients with subacromial pain syndrome the activity of the serratus anterior muscle (SA) decreases during scaption exercises with different weight loads compared to the healthy population.^11,12 In contrast, collegiate baseball players have demonstrated decreased activity of the UT on the dominant side during both GHJ flexion and abduction exercises compared with the non-dominant side, with increased activity of the LT on the dominant side.¹³ The LT muscle attaches to the spine of the scapula and plays a critical role in the posterior tilt of the scapula during upward rotation while producing a force couple with other scapular muscles, which helps to maintain subacromial space width.¹⁴

Hyperactivity of the deltoid has been suggested to reduce the amount of subacromial space width as a result of superiorly directed humeral head translation during arm elevation relative to the glenoid fossa.^15-17 Thus, clinicians need to be aware of deltoid activity during shoulder and scapular exercises to avoid over recruitment of the deltoid muscle, especially when working with overhead athletes. Additionally, the appropriate intensity for external rotation isometric exercise has been examined with respect to the EMG activity of the deltoid and infraspinatus by Bitter et al, who concluded that 40% of maximal voluntary isometric contraction in the infraspinatus muscle was optimal.¹⁵ Moreover, the scapular plane has been suggested as the position to maintain optimal bony congruity between the humeral head and glenoid fossa as well as the optimal length-tension relationship of the scapulohumeral musculature.¹⁸ Modulation of rotator cuff muscle activation has been observed across different weight loads at two different angular velocities during scaption exercises while measuring deltoid muscle activity.¹⁹ No study, however, has investigated the activity of the LT and SA muscle relative to the activity of the anterior deltoid (AD) muscle during scaption exercises. Therefore, the purpose of this study was to investigate the level of LT and SA muscle activity during scapular plane elevation (scaption) in three positions while maintaining a moderate level of AD muscle activity. A secondary purpose was to identify the ratio of UT to LT muscle activity during the varied scaption exercises. The authors hypothesized that the activation of these two important muscles and the UT/LT ratio would vary with exercise position and throughout the range of scapular plane elevation.

METHODS

Participants

Fourteen active healthy male collegiate subjects belonging to the university badminton club (age: 20.3 ± 2.2 years, height: 172 ± 6.0 cm, weight: 66 ±  6.9 kg, competitive experience: 6.3 ± 1.8 years) agreed to participate in this study. All subjects gave informed consent to the procedures as approved by the Institutional Review Board of the University (IRB protocol: F1304019) prior to the examination. All subjects indicated no history of injuries, neurologic or physiologic deficits in the upper body on a preliminary screening questionnaire. The subjects were instructed to wear their own athletic shoes during the examination which required approximately 45 minutes to complete.

Experimental procedure

The subjects elevated the dominant arm with their elbow extended in the scapular plane (scaption), from the side of the body to 180 degrees of GHJ abduction or as much abduction as possible, in three different positions: 1) standing (STAN), 2) quadruped (QUAD), and 3) prone (PRON). The subjects performed the scaption exercises with three different weight loads (dumbbells) for each of the exercise positions: 0 kg (no weight) 1.8 kg and 4.1 kg. Those three weight loads were determined based on the results of a previous study.²⁰ The subject performed the standing scaption exercise from the lateral side of the hip to full abduction (Figure 1a). The subject was instructed to avoid leaning the upper body backward. The subject was also instructed to maintain the arm in a position such that the thumb was pointing upward toward the ceiling up to 90 °, after which they were instructed to point the thumb backward. The QUAD position was performed on a treatment table, the subjects started to move the arm with the hand at the level of the hip (Figure 1b), whereas in the PRON position the subjects lay down at the edge of treatment table and started to move the arm from the hand placed on the table (Figure 1c). Prior to scaption exercises, the subjects were instructed on how to adjust the arm to 40 ° (in order to be in the scapular plane) as measured by a goniometer, and learned the angle for all three testing positions. The subjects performed scaption exercises without volitional retraction or protraction of the scapula for three repetitions for each of the three weight loads in each of the three positions. The weight loads were progressively increased from 0 kg to 4.1 kg for all of the subjects. The movement speed for each of the exercises was also controlled by a metronome set at a frequency of 1 Hz or 1 beat per second (sec). Accordingly, the subjects were asked to move the arm smoothly throughout the range of motion from the side of the body to as much abduction in the scapular plane as possible for 5 sec with a steady and constant tempo.

Figure 1.

The three seconds positions of abduction exercise in the scapular plane (scaption) with a 1.8 kg dumbbell, a) the standing position, b) the quadruped position, and c) the prone position.

Data Management and Analyses

Surface electromyography (EMG) was utilized to measure three scapular muscles on the dominant side: the LT, the SA, and the UT. Surface EMG was also utilized to measure the AD and PD muscles, which enabled the identification of the level of muscle activity during the movement for these muscles. The skin surface was prepared by vigorously cleaning with an alcohol swab to minimize skin impedance before electrode placement. This study used bipolar surface silver EMG electrodes (model Delsys Bagnoli-8; Delsys Inc, Natick, MA) with a bar length of 10 mm, a width of 1 mm, and a distance of 1 cm between active recording sites. The electrodes were placed at an oblique angle from the scapular spine and just outside the medial border of the scapula for the LT muscle; below the axilla between the latissimus dorsi and pectoralis major at the level of the scapular inferior angle for the SA muscle; and at halfway between the C7 spinous process and the acromion process for the UT muscle.²¹ Also, the electrodes were placed 2 cm inferior to the lateral border of the clavicle and angled parallel to the muscle fibers for the AD muscle, and at an oblique direction parallel to the muscle fibers of the deltoid muscles at the lateral border of the scapular spine for the PD.⁸ The reference electrode was placed between the LT and PD electrodes.

Once the electrodes were secured using tape, the maximum voluntary isometric contraction (MVIC) for each of the muscles was measured by using the manual muscle-testing procedures for the normalization of EMG data. The root-mean-square (RMS) values of the EMG signals for the UT, SA, and AD were normalized to the MVIC of the corresponding muscles during scapular plane elevation at 90 ° of GHJ in a standing position,⁸ whereas the EMG signal for the LT was normalized to the MVIC of the corresponding muscle at 180 ° of GH flexion or with as much flexion as possible in a quadruped position in which the hips and knees were flexed at 90 °.⁸ Also, the EMG signal for the PD was normalized to the MVIC of the corresponding muscle at 90 ° of GH horizontal abduction.

Input signals of EMG activities were recorded using a data collection system (MP 150 Data Acquisition System; BIOPAC System Inc, Goleta, CA, USA) with a sampling rate of 1000 Hz, and all data was stored in a hard drive for off-line analyses. The EMG electrodes were preamplified (10×) and routed through the EMG mainframe, which further amplified (100×), a total gain of 1000 × and band-pass filtered (20-450 Hz) signals. The RMS for the LT, SA, UT, AD, and PD were normalized to the MVIC of the corresponding muscles as described above for further analyses.

Each of the data sets consisted of 5000 samples of RMS activity as a dependent variable measured for 5 sec from initial muscle activity to the completion of each exercise (1000 Hz × 5 sec). This study analyzed the dependent variables of the LT, SA, and UT, which were simultaneously blocked at every 1000 samples of the RMS activity or every 1 sec during each of the exercises.²⁰ Consequently, 5000 samples of RMS activity for each exercise were divided into five blocks of dependent variables from the initial EMG activity to the end of the exercise.

For data analyses of the normalized RMS activity of the muscle, a 3 × 3 × 5 (position × weight load × block) repeated-measures analysis of variance (ANOVA) design within subjects crossed with positions, weight loads and blocks was used to identify differences in each mean value of the normalized RMS activities of the LT and SA muscles. Also, a 3 x 3 x 5 repeated-measures ANOVA design was used to identify differences in ratios of the UT to LT muscle activities. Where appropriate, the simple main effect and Tukey's honestly significant different post hoc test (Tukey's HSD) were used to identify any significant difference for each normalized RMS activity. This study further used Pearson correlation coefficients to determine if there was a relationship across EMG activities. All statistical tests were performed at the 0.05 level of probability (p < .05).

RESULTS

Lower Trapezius and Serratus Anterior

A within-subject (subject 3 trial) ANOVA design was used to calculate the ICCs. The highest mean of the ICCs (2, 1) as tested by the same tester was 0.74 of an individual's true score, which showed the consistency of the measure, for LT muscle activity with the weight load of 4.1 kg. It was also 0.75 for the QUAD position. The highest mean of the ICCs (2, 1) was 0.75 for SA muscle activity with 4.1 kg. It was also 0.71 for the STAN position. Each of the ICCs is presented in the Table 1.

Table 1.

Intraclass Correlation Coefficients for Reliability of Lower Trapezius and Serratus Anterior Muscle EMG Activities, across the three repetitions for all positions, loads, and across blocks.

Lower Trapezius
	STANDING			QUADRUPED			PRONE
BLOCK (sec)	0 kg	1.8 kg	4.1 kg	0 kg	1.8 kg	4.1 kg	0 kg	1.8 kg	4.1 kg
1	0.69	0.91	0.71	0.58	0.85	0.88	0.29	0.70	0.87
2	0.59	0.78	0.80	0.63	0.73	0.89	0.60	0.62	0.74
3	0.48	0.63	0.75	0.52	0.75	0.86	0.43	0.42	0.67
4	0.83	0.61	0.41	0.76	0.73	0.79	0.50	0.62	0.55
5	0.61	0.62	0.72	0.91	0.49	0.88	0.68	0.71	0.57

Serratus Anterior
	STANDING			QUADRUPED			PRONE
BLOCK (sec)	0 kg	1.8 kg	4.1 kg	0 kg	1.8 kg	4.1 kg	0 kg	1.8 kg	4.1 kg
1	0.73	0.51	0.49	0.33	0.82	0.83	0.19	0.70	0.67
2	0.23	0.70	0.87	0.40	0.59	0.87	0.32	0.58	0.80
3	0.84	0.84	0.82	0.43	0.38	0.92	0.28	0.56	0.53
4	0.70	0.60	0.72	0.63	0.74	0.67	0.48	0.71	0.73
5	0.92	0.90	0.77	0.85	0.87	0.58	0.45	0.89	0.91

Numbers 1-5 = the time period, in seconds, from the initial activity to the completion of the performed movement.

Mean values and 95% confidence intervals for LT EMG activities are presented in Table 2 and Figure 2. The mean values of LT EMG activity in STAN were significantly lower at all of the range of motion blocks with the weight loads of 1.8 kg and 4.1 kg than those of both of QUAD and PRON, except for the mean value at the block of 5 sec with the weight load of 1.8 kg [the critical value of the Tukey HSD (D_Tukey) = 7.86%, p < .05]. No difference was observed in the mean values between QUAD and PRON, except for the mean values at the block of 2 sec with the weight load of 4.1 kg in which the mean value was significantly greater in QUAD than that of PRON (40.8% and 30.2% of MVIC, respectively).

Table 2.

Mean EMG activity of Lower Trapezius, reported as % MVIC (95% CI's).

STANDING	1	2	3	4	5
0 kg	5.4 (3.4, 7.4)†‡‡	7.4 (4.6, 10.1)††‡‡	10.9 (7.9, 14.0)*††‡	20.1 (14.8, 25.4)*††‡‡‡	36.7 (24.2, 49.2)††‡‡‡‡
1.8 kg	9.2 (6.3, 12.1)*‡‡‡‡	16.7 (12.8, 20.6)*††‡‡	22.2 (18.7, 25.7)**††‡‡	32.4 (26.9, 37.9)**††‡‡	45.6 (33.8, 57.4)†‡‡‡
4.1 kg	10.6 (6.6, 14.7)*†‡‡‡‡	25.4 (17.4, 33.3)*††‡‡‡‡	37.0 (29.9, 44.1)**††‡‡‡	44.5 (39.1, 49.9)**††‡‡	50.7 (41.5, 59.8)**†‡‡

QUADRUPED	1	2	3	4	5
0 kg	11.0 (8.7, 13.2)††‡‡	15.9 (13.5, 18.3)††‡‡	22.3 (18.9, 25.7)*††‡	32.2 (23.0, 41.4)*††‡‡	39.2 (25.9, 52.5)††‡‡‡
1.8 kg	16.6 (11.7, 21.4)**††‡‡‡	25.6 (20.9, 30.3)*††‡‡‡	39.0 (33.2, 44.8)*††‡‡‡	48.5 (40.0, 57.1)*††‡‡	52.3 (39.9, 64.6)††‡‡‡
4.1 kg	25.7 (16.7, 34.6)*††‡‡‡‡	40.8 (31.3, 50.4)**††‡‡	55.5 (44.7, 66.4)*††‡‡	62.3 (52.9, 71.7)*††‡‡	63.2 (51.1, 75.4)*††‡

PRONE	1	2	3	4	5
0 kg	4.3 (2.4, 6.2)†‡‡‡	8.4 (6.0, 10.8)††‡‡	16.4 (13.3, 19.5)††‡‡‡	31.8 (22.5, 41.1)*††‡‡	37.7 (26.5, 48.8)††‡‡‡
1.8 kg	8.3 (6.0, 10.7)‡‡‡	17.5 (13.1, 21.9)††‡‡‡	33.5 (26.6, 40.4)*††‡‡‡‡	47.0 (37.3, 56.6)*††‡‡‡	51.9 (39.7, 64.1)††‡‡‡
4.1 kg	15.9 (9.9, 21.9)†‡‡‡‡	30.2 (21.2, 39.3)*††‡‡‡‡	49.4 (39.3, 59.5)*††‡‡‡‡	61.7 (50.8, 72.5)*††‡‡‡	62.9 (50.5, 75..4)*††‡‡‡

Numbers 1-5 = the time period, in seconds, from the initial activity to the completion of the performed movement.

* indicates a significant difference across the different positions at each block and each weight load [the critical value of the Tukey HSD (D_Tukey) = 7.86%, p < 0.05]. † indicates a significant difference across the different weight loads in each position at each block (D_Tukey = 7.85%, p < 0.05). ‡ indicates a significant difference across the different blocks at each weight load in each position (D_Tukey = 11.4%, p < 0.05).

Figure 2.

The means of root-mean-square (RMS)-electromyography (EMG) activities for the lower trapezius muscle [the percent of maximum voluntary isometric contraction (% MVIC)] are shown in three different positions: standing, quadruped, and prone. Dash lines with black squares indicate the weight load of 0 kg; dash lines with white circles indicate the weight load of 1.8 kg; and dot lines with white triangles indicate the weight load of 4.1 kg. The error bars denote the standard error of the mean.

The mean values of LT EMG activity were significantly increased at all of the blocks in QUAD while the weight loads were progressively increased for each of the positions (D_Tukey = 7.85, p < .05). Last, the mean values were significantly greater at the blocks of 4 sec and 5 sec than those of the blocks of 1 sec and 2 sec in all of the positions (D_Tukey = 11.4, p < .05), whereas no difference was observed between the block of 4 sec and 5 sec.

Mean values and 95% confidence intervals for SA EMG activities (% MVIC) are presented in Table 3 and Figure 3. The mean value of SA EMG activities in STAN was significantly greater at all of the blocks with the weight loads of 1.8 kg and 4.1 kg than those of both of QUAD and PRON, except of the mean values at the block of 5 sec [D_Tukey = 8.50%, p < .05]. Also, a significant difference in mean values was observed between QUAD and PRON at the block of 4 sec with both weight loads of 1.8 and 4.1 kg (29.7% and 52.0% of MVIC for QUAD with the weight load of 1.8 kg vs. 20.4% and 37.7% of MVIC for PRON; 41.6% and 65.6% of MVIC for QUAD with the weight load of 4.1 kg vs. 30.7% and 54.6% of MVIC, respectively) (p < .05). Finally, the mean value was significantly greater in QUAD at the block of 3 sec with the weight load of 4.1 kg than that of PRON (16.2% and 9.5% of MVIC respectively) (p < .05).

Table 3.

Mean EMG activity of Serratus Anterior reported as % MVIC (95% CI's).

STANDING	1	2	3	4	5
0 kg	10.5 (7.8, 13.1)†‡‡‡	18.3 (14.6, 22.0)**††‡‡	23.9 (21.0, 26.7)**††‡	28.6 (24.5, 32.7)*††‡‡	32.8 (24.6, 41.0)††‡‡
1.8 kg	15.5 (11.7, 19.3)**‡‡‡‡	29.1 (23.6, 34.6)**††‡‡‡‡	42.9 (37.0, 48.8)**††‡‡‡‡	52.8 (45.4, 60.2)**††‡‡‡	54.5 (44.7, 64.3)*††‡‡‡
4.1 kg	18.9 (13.2, 24.6)**†‡‡‡‡	41.7 (32.3, 51.1)**††‡‡‡‡	65.7 (57.2, 74.2)**††‡‡‡‡	80.3 (72.4, 88.2)**††‡‡‡	72.0 (59.8, 84.2)*††‡‡‡

QUADRUPED	1	2	3	4	5
0 kg	3.0 (2.2, 3.7)‡‡	4.0 (2.8, 5.2)*‡‡	9.3 (7.0, 11.7)*†‡‡	20.2 (17.1, 23.2)††‡‡‡‡	34.1 (24.9, 43.4)††‡‡‡‡
1.8 kg	4.8 (3.4, 6.3)*‡‡	4.7 (3.4, 5.9)*‡‡	11.1 (8.9, 13.3)*‡‡	29.7 (24.8, 34.6)**††‡‡‡‡	52.0 (39.5, 64.5)*††‡‡‡‡
4.1 kg	5.9 (4.0, 7.7)*‡‡‡	5.6 (4.2, 7.1)*‡‡‡	16.2 (12.7, 19.6)*†‡‡‡‡	41.6 (33.7, 49.4)**††‡‡‡‡	65.6 (53.6, 77.6)*††‡

PRONE	1	2	3	4	5
0 kg	3.1 (1.6, 4.5)†‡‡	2.9 (2.2, 3.7)*‡‡	7.0 (4.7, 9.2)*‡‡	17.5 (12.3, 22.7)*†‡‡‡‡	27.0 (20.3, 33.8)††‡‡‡‡
1.8 kg	4.8 (3.2, 6.5)*‡‡	3.7 (2.4, 4.9)*‡‡	6.9 (5.2, 8.7)*‡‡	20.4 (14.6, 26.1)**†‡‡‡‡	37.7 (28.9, 46.4)**††‡‡‡‡
4.1 kg	9.5 (5.3, 13.7)†‡‡	6.7 (4.1, 9.2)*‡‡	9.5 (7.0, 11.9)‡‡	30.7 (22.4, 38.9)**††‡‡‡‡	54.6 (43.5, 65.7)**†‡‡‡

Numbers 1-5 = the time period, in seconds, from the initial activity to the completion of the performed movement.

* indicates a significant difference across the different positions at each block and each weight load [the critical value of the Tukey HSD (D_Tukey) = 7.86%, p < 0.05]. † indicates a significant difference across the different weight loads in each position at each block (D_Tukey = 7.85%, p < 0.05). ‡ indicates a significant difference across the different blocks at each weight load in each position (D_Tukey = 11.4%, p < 0.05).

Figure 3.

The means of root-mean-square (RMS)-electromyography (EMG) activities for the serratus anterior muscle [reported as % MVIC) are shown in the three different positions: standing, quadruped, and prone. Dashed lines with black squares indicate the weight load of 0 kg; dashed lines with white circles indicate the weight load of 1.8 kg; and dotted lines with white triangles indicate the weight load of 4.1 kg. The error bars denote the standard error of the mean.

The mean values of SA EMG activity were significantly increased at all of the blocks in STAN while the weight loads were progressively increased (D_Tukey = 6.17, P < .05). However, the mean values were not increased at all of the blocks in both of QUAD and PRON until the block of 4 sec. Finally, the mean values were significantly increased in STAN up to the blocks of 4 sec (D_Tukey = 9.23, p < .05). In contrast, the mean values were significantly greater in both of the QUAD and PRON at the blocks of both of 4 sec and 5 sec than those of the blocks from 1 sec to 3 sec.

Ratio of Upper Trapezius to Lower Trapezius

The UT/LT ratio decreased as the arm was elevated during the STAN position, while the UT/LT ratio increased as the arm was abducted in the scapular plane during the QUAD position. Specifically, the UT/LT ratio in STAN was significantly greater at the blocks between 1 and 3 sec than that during both QUAD and PRON (1.95, 1.85, and 1.69 for STAN; .49, .53, and .69 for QUAD; and .89, .76, and .86 for PRON, respectively) (D_Tukey = .33, p < .05). The UT/LT ratio in STAN was also significantly greater than that of QUAD at the block of 4 sec (1.41 and .95, respectively) (p < .05). No difference in the UT/LT ratios was observed between QUAD and PRON, regardless of blocks.

With regard to a comparison of different blocks, the UT/LT ratios were significantly greater at both blocks of 1 sec and 2 sec than those of blocks of 4 and 5 sec for STAN (1.95, 1.85, 1.41 and 1.10, respectively) (D_Tukey = .35, p < .05). Also, the ratio for STAN was significantly greater at the block of 3 sec than that of the block of 5 sec (1.31 and 1.10, respectively). In contrast, for QUAD, the UT/LT ratios at both blocks of 1 and 2 sec were significantly less than those of both blocks of 4 and 5 sec (.49, .53, .95, and 1.30, respectively) (p < .05). Likewise, in PRON, the UT/LT ratios were significantly less at the blocks of 1 to 3 sec than that of the block of 5 sec (.89, .76, .86, and 1.34, respectively) (p < .05) (Figure 4).

Figure 4.

The means of the ratio of upper trapezius muscle activity to lower trapezius muscle activity are shown in three different positions: standing, quadruped, and prone. The error bars denote the standard error of the mean.

Anterior Deltoid and Posterior Deltoid

Mean values for AD and PD EMG activities (% MVIC) are presented in Table 4. It was observed that STAN scaption exercises activated the AD muscle more than the PD muscle with each of the weight loads across different blocks while QUAD and PRON exercises activated the PD muscle more than the AD muscle. It is plausible that the high effect of PD in both QUAD and PRON exercises was due to the effect of gravity. It is interesting to note that the correlation coefficient in the amount of EMG activity between the SA and AD muscle was .82 (p < .001), regardless of the position, weight loads, and blocks, whereas the correlation coefficient between the SA and PD muscle was .05. It is plausible to suggest that the SA and AD muscles were activated as the manner of synergy activation in contractions.

Table 4.

Mean EMG activity of Anterior and Posterior Deltoids reported as % MVIC, across positions and across the 5 blocks.

% MVIC	STANDING					QUADRUPED					PRONE
ANT DELTOID	1	2	3	4	5	1	2	3	4	5	1	2	3	4	5
0 kg	12	19	23	28	39	3	5	10	23	39	2	4	9	20	32
1.8 kg	17	28	39	44	47	6	10	16	30	50	4	7	13	27	40
4.1 kg	18	43	65	71	63	10	18	28	45	58	10	16	24	40	55
POST DELTOID
0 kg	3	5	8	13	20	10	15	19	22	22	11	16	21	25	25
1.8 kg	4	9	16	22	26	26	37	41	42	35	30	41	49	48	37
4.1 kg	6	17	30	36	36	47	61	64	55	43	58	68	71	65	53

Numbers 1-5 = the time period, in seconds, from the initial activity to the completion of the performed movement.

DISCUSSION

Quadruped Scaption Exercise

Exercise intensity can be determined by the guideline generalized by DiGiovine et al. (1992)²¹ in which a range of 0% to 20% MVIC was considered low activity, 21% to 40% for moderate activity, 41% to 60% for high activity, and greater than 60% for very high activity. With this concept, this study determined the maximum amount of LT muscle activity occurred when there was a moderate level (40% MVIC or less) of AD muscle activity. Accordingly, QUAD scaption with the weight load of 1.8 kg at 4 sec after the initial movement in the scapular plane activated the LT muscle up to 49%, which was normalized to the MVIC of the corresponding muscle (MVIC), followed by PRON scaption with the same weight load, which activated up to 47% MVIC. These amounts were significantly greater than activation of LT during STAN scaption exercise (32% MVIC). The arm elevated at 120 ° of GHJ abduction has been suggested to be aligned with the muscle fiber orientation of the lower trapezius.²³ Oyama et al⁷ examined six scapular-retraction exercises with a variety of shoulder angles in the prone position, including 120 ° GHJ abduction. These authors demonstrated that prone 120 ° GHJ abduction exercise activated the LT muscle up to 68% MVIC without weight loads while the participants were instructed to “squeeze the shoulder blades together.”⁷ The prone 120 ° GHJ abduction exercise with scapular retraction, however, increased UT muscle activity up to 72% MVIC.⁷ This current study instead revealed that the activity of UT muscle was 41% for QUAD and 40% for PRON, which was defined as a moderate level of muscle activity. Thus, QUAD and PRON scaption exercises could effectively minimize UT muscle activity, compared with Oyama's results. In addition, unlike middle deltoid activity, the hyperactivity of the PD muscle has not been evident in the adverse effect on the subacromial space width as a result of humeral head translation.^16,17,19

Standing Scaption Exercise

STAN scaption exercise with the weight load of 1.8 kg at 3 sec after the initial movement in the scapular plane still produced moderate levels of AD muscle activity (39% MVIC) while generating 22% MVIC of the LT muscle and 43% MVIC of the SA muscle. Alpert et al¹⁹ examined standing scaption exercises with two different angular velocities and five different weight loads: 0%, 25%, 50%, 75%, and 90% of normalized maximum weight (NMW), whose averages ranged between 2.3 and 8.4 kg. These authors found that the nearly 100% or even more than 100% MVIC of AD EMG activity was observed between 60 ° and 90 ° of GHJ in the scaption exercise with 25% and 50% NMW at the angular velocity of 100 °/sec or 1.75 rad¹⁸ This angular velocity was three time faster than the angular velocity used in the current study that required the subjects to bring the arms to the maximum elevation over 5 sec, which is equivalent to 36 °/sec (180 °/5 sec) or .63 rad. Thus, the amount of AD EMG activity seems to vary with angular velocity or movement speed.¹⁹ Clinically, it is suggested that scaption exercises be performed more slowly in order to minimize AD EMG activity. Tsuruike et al.²¹ found that the mean value of AD muscle activity was 47% MVIC during GHJ flexion exercise with the weight load of 1.8 kg and 43% MVIC in GHJ abduction exercise at 3 sec after the initial movement. Consequently, it appears that scaption exercise in the standing position relatively minimizes AD muscle activity compared with GHJ flexion and abduction. Recently, exercise at 30 ° of GHJ abduction, flexion, and 90 ° of elbow flexion with the thumb up, called the “champagne toast position,” has been demonstrated to minimize deltoid muscle activity (38% of maximal manual testing: MMT), compared with the traditional “full can” test position at 90 ° of GHJ abduction (60% MMT).²⁴ Consequently, the “champagne toast” needs to be compared with the results of the current study.

UT/LT Ratio

The ratio of UT to LT is important to consider during shoulder rehabilitation exercise. Particularly, the activity of the LT muscle should be emphasized in the early stage of rehabilitation to improve scapular kinesis while minimizing UT muscle activity.^9,14 This study identified significantly high UT/LT ratios in STAN scaption exercise up to the block of 4 sec compared with those of QUAD exercise. STAN scaption with a weight load of 1.8 kg could be performed up to 90 ° of GHJ elevation while the exercise maintained a moderate level of AD muscle activity. Additionally, the results of the current study indicate that QUAD scaption exercise with a weight load of 4.1 kg can be performed up to 4 sec after the initial movement to enhance SA muscle activity (42% MVIC) at slightly more than the moderate levels of AD muscle activity (45% MVIC).

The LT and SA muscles have been identified as fatigue-resistant muscle fibers, compared with the rotator cuff muscles.²⁵ With this, the number of repetitions in exercise can be more important than the amount of weight loads, which readily activate AD and UT muscle. Scapular exercises which activate the LT and SA muscles have been demonstrated in a number of previous studies for sport-related specific rehabilitation.^{2,4-8,20,26-29} However, the AD muscle, whose hyperactivity possibly superiorly translates the humeral head, leading to a decrease in subacromial space width. Also, the exercises with high ratios of UT/LT may deteriorate the scapulohumeral rhythm, which may be associated with individuals with scapular dyskinesis. Those two concerns must be addressed in rehabilitation programs especially for overhead athletes. Further studies are warranted to investigate LT and SA muscle activity in the application of rotator cuff and scapular exercise, such as horizontal abduction with the elbow flexed in the quadruped position against gravity.

Limitations

This study included a sample delimited to active male collegiate participants in the badminton club with habitual adaptations and a sports specific arm dominance. Thus, the participant cohort may limit the generalization of these findings to injured populations. Also, the methods used in the study did not allow identification of the range of motion while using the metronome that consistently controlled the pace of the arm movement. Thus, exactly where peak activity occurred in the range of motion was not able to be described. This study identified LT and SA muscle activities during scaption exercises while the AD maintained muscle moderate activity. However, the authors are unable to identify if there would be any compensation or changes that could occur with greater levels of the AD muscle. Finally, this study did not measure rotator cuff muscle activation, which may be a further limitation and be a suggestion for inclusion in future studies.

CONCLUSION

The results of the current study indicate that both LT and SA muscle activation was dependent upon the relative movement of the arm into elevation during the varied scaption exercises. Up to 63% MVIC of LT muscle activity was observed in QUAD scaption exercise, whereas up to 80% MVIC of SA muscle activity was associated with STAN scaption exercise. However, based on UT/LT ratios QUAD scaption exercise could be recommended with weight load up to 4.1 kg to improve both LT and SA muscle activation in young active overhead athletes with asymptomatic shoulders.